Jemly Privacy Policy

Effective Date: 2026-11-11
Last Updated: 2026-11-02

The Service operator (“we” or “us”) explains in this Privacy Policy (the “Policy”) how we collect, use, store, share, and delete personal information when you use the Jemly iOS app and related services (the “Service”), and what rights you have.

By tapping Agree, registering, or using the Service, you acknowledge this Policy. For sensitive uses (e.g., Style Help profile upload) or optional permissions, we will seek separate consent where required. Together with the User Agreement, this Policy governs the Service; for personal data, this Policy controls.

Your privacy rights depend on the laws applicable to you, including the GDPR / UK GDPR and U.S. state privacy laws where applicable. Section 11 explains how to exercise these rights.

Contents

1. Who we are and scope
2. Data storage location
3. Personal information we collect
4. Profile image and body metrics
5. AI processing (memory forward, providers, no training, audits)
6. Purposes and legal bases
7. Device permissions, SDKs, sign-in, and payments
8. Personal information sharing, processors, and disclosure
9. International transfers
10. Retention, deletion, and account closure
11. Your privacy rights
12. Security
13. Children
14. Policy updates
Appendix: Processors / third parties and SDKs
15. Contact us

1. Who we are and scope

1.1 The Service operator is the controller. Operator and contact details appear at the end of this Policy.

1.2 This Policy covers the Jemly App. It does not cover Apple, Google, or other third parties’ own services (their policies apply on their pages).

1.3 Authorized personnel may access relevant information only as needed to provide the Service, address security incidents, investigate faults or respond to your requests. Access is controlled and logged.

2. Data storage location

Your account, wardrobe, credits and audit data are stored and processed in the United States, with processing by the service providers described below.

3. Personal information we collect

We collect only what is needed for the features you use. Declining optional data affects only related features.

3.1 Information you provide

3.2 Information collected automatically

3.3 Information from third parties

• Apple / Google: account identifiers and data you authorize when using their sign-in.
• Apple payments: payment result and necessary transaction data; we do not collect full card numbers.
• Open-Meteo: we send city or approximate coordinates for weather; we do not attach your name or plaintext account id beyond what is needed to run the query.

3.4 What we do not routinely collect

We do not require government ID for ordinary registration. We do not read your contacts. We do not use the microphone. We do not use continuous precise location tracking as a default feature. We have no advertising SDK. Firebase Analytics and Firebase Crashlytics are integrated for technical diagnostics as described in Section 6.1 and the Appendix. These diagnostics are not used to analyze private content, infer personal preferences, or build advertising profiles.

4. Profile image and body metrics

4.1 Higher-sensitivity data

Profile photos that show your face or body, and height/weight, can be sensitive. We apply minimization, separate consent where required, and stronger safeguards.

4.2 On-device by default

• Profile photo and height/weight are stored in the iOS app sandbox (per-account local DB) by default and are not routinely synchronized as profile records. Height/weight used for a generation may be retained in the outfit record context and the complete audit prompt under Section 5.5.
• For try-on / recommendation generation, data is sent over encrypted transport to our API and memory-forwarded to the model provider for that request. This forwarding does not separately persist original input images in queues, logs or temporary files. Wardrobe storage, consented Style Help sharing, delivery caches and generated audit images follow their distinct retention rules.

4.3 Separately consented sharing: Style Help

Only after you separately check consent to publish a Style Help invite do we upload a profile snapshot and selected garment images to our servers for the invite window.

• The invitation follows the validity period shown and confirmed by you when it is published.
• On expiry or when you delete the invite/results, access stops and cleanup jobs run.
• Refusing consent does not block wardrobe browsing or text suggestions that do not need the upload; you simply cannot publish invites that require the snapshot.

Authorized moderation staff may view uploaded invitation materials (profile snapshots and garment images) and generated Style Help results on a need-to-know basis for content safety, complaints and disputes. We record relevant invitation or result identifiers, review status, processing times and necessary rejection reasons for content safety, handling disputes and necessary traceability. Access to these records is permission-controlled and retention follows the minimum-necessary principle in Section 10. Images follow the retention and deletion rules for the relevant materials and generated results; moderation does not extend image retention. This review does not access unuploaded photos on your device or grant any model-training permission.

4.4 What we do not do

• No face login, 1:1 / 1:N identity matching, or cross-user face search;
• No selling profile photos for ad profiling;
• No model training on this data without your separate explicit consent (Section 5.4).

4.5 Third-party likeness

Upload only photos you have rights to use. Rights holders may contact the contact email at the end of this document.

5. AI processing

Core rules:

1. Avatar (profile photo) and height/weight stay on-device by default; selected images and necessary parameters are transmitted for AI processing, and measurements used for a generation may be retained in its outfit context and audit prompt; Style Help separately stores a time-limited snapshot after your consent.
2. Generated outfit / try-on images are saved on your device after delivery; we do not run a cloud cross-device gallery of those results.
3. Audit images and metadata are retained for up to 6 months from their respective creation or archival, automatically cleared on a recurring schedule, and deleted earlier when you close your account. Audit records include complete prompts and any height/weight used for that generation. Access is limited to internal safety, anti-abuse, troubleshooting and dispute handling, subject to lawful disclosure.
4. We do not use your inputs or Outputs for model training without separate explicit consent. Third-party processing and retention follow the arrangements described in this Policy and applicable service terms.

5.1 What we send and how

When you use garment localization, recognition, text styling, try-on or recommendation image features, we may send the relevant garment images or attributes, profile photo, height/weight where needed, and your styling requirements and prompts after necessary safety checks. Processing queues carry task identifiers, not image files.

Memory forward: original AI input images are forwarded in memory without separately persisting them in request queues, logs or temporary files. Wardrobe photos, consented Style Help materials, delivery caches and generated audit images are stored under their respective rules in this Policy.

5.2 Model providers

Recipients, purposes and information transmitted are detailed in Section 8.1, Personal information sharing. API keys stay in server-side secure configuration, never in the client.

Providers may perform short-lived processing or technical retention under their terms and DPAs (e.g., abuse monitoring, debugging). Retention follows executed contracts/configs. Our no-training commitment is not a promise of zero retention by third parties. Before transferring personal data, we request separate AI processing consent; provider changes require renewed disclosure and consent. Declining or withdrawing stops new transfers, but cannot undo processing already started.

5.3 Storing Outputs

After delivery confirmation, try-on / recommendation images are saved on your device. We do not persist generated images as a user content library for cross-device viewing. Personal generation delivery caches last up to 7 days and are removed sooner after acknowledgement. Generated audit copies are retained separately for up to 6 months and deleted earlier on account closure (Section 5.5). Style Help shared results follow the period displayed for that feature. You may delete local images; uninstalling the App or clearing App data removes them—back up first.

5.4 No training (core commitment)

We use your inputs and Outputs only to fulfill your requests, provide and maintain the Service, and protect security and compliance. Without your separate explicit consent, we will not use your inputs (including profile photos, garment images, height/weight, prompts) or Outputs to train, fine-tune, or improve machine-learning models.

5.5 Generation audits

For internal safety, anti-abuse, troubleshooting and billing or delivery disputes, we record account and task identifiers, generation type, input-image and garment identifiers, model, status, duration, credit snapshots, provider error details and usage, and the complete prompt sent for image generation. Prompts may include the height/weight and clothing descriptions used for that generation. Input audit records reference identifiers rather than storing original input images.

Audit metadata and generated audit images are retained for up to 6 months from their respective creation or archival and automatically cleared on a recurring schedule. Account closure automatically deletes associated audit records and clears audit images and stored copies without waiting for that period to end. Failed image deletions are retried. Audits are not a user cloud gallery or a public resource; access is controlled and logged.

Authorized personnel may inspect relevant audit images and records as necessary to investigate complaints, security incidents or disputes. Internal hold markers do not automatically extend the stated retention period or prevent deletion on account closure.

5.6 AI-generated content notice

Outputs may be disclosed or labeled as AI-generated for transparency or platform rules.

6. Purposes and legal bases

6.1 Purposes

1. Create and maintain accounts, login, and deletion;
2. Provide wardrobe, profile, recommendations, try-on, Style Help;
3. Credit reserve/settle, top-ups, redemptions, billing;
4. Security, fraud/abuse detection, troubleshooting;
5. Customer support;
6. Legal obligations and lawful requests from authorities;
7. Other uses with your consent.

App performance, stability, and crash diagnostics. We use Google Analytics for Firebase (Firebase Analytics) and Firebase Crashlytics to identify slow or unsuccessful operations, assess the impact of failures, analyze crash causes, and verify fixes. Our analysis concerns how the software runs; it is not an analysis of your identity, private photos or messages, body characteristics, or clothing preferences, and is not used for advertising profiles or model training.

Diagnostic events can include predefined screen names and operation categories, camera/library import source, recognition method, item counts, elapsed time, success/failure/cancellation classifications, and whether a generated result became available, was saved locally, or was displayed. These technical events help locate the stage at which an operation becomes slow or fails. We do not attach account IDs, names, phone numbers, email addresses, photos, image URLs, garment names or brands, prompts, body measurements, precise location, task IDs, or raw business-error messages to these events. We do not set an account-linked user ID or custom user profile in these SDKs.

The SDKs also process technical information such as app/device/OS versions, app-instance or installation and session identifiers, and crash stack traces. Network requests necessarily expose connection information such as an IP address to the recipient. These identifiers support diagnostic deduplication and session statistics; this is not completely anonymous or identifier-free processing. The integration does not include IDFA support; IDFV collection and advertising-personalization signals are disabled.

The SDKs are currently enabled in development/test builds, while collection remains disabled in the production release configuration. Before production collection is enabled, we will provide the applicable notice and obtain consent or complete other required procedures. This disclosure does not itself enable collection or substitute for any required consent.

6.2 Legal bases (GDPR / UK GDPR, where applicable)

Where we rely on legitimate interests, you may object (Section 11). Where we rely on consent, you may withdraw it without affecting prior lawful processing.

6.3 U.S. notice (general)

We collect the categories in Section 3 for the business purposes in Section 6.1. We do not use sensitive personal information for inferring characteristics beyond providing the features you request (try-on / Style Help). See Section 11 for California rights.

7. Device permissions, SDKs, sign-in, and payments

7.1 iOS permissions

Permissions are not on by default; the system prompt is yours to accept or deny.

We do not request Contacts access. We do not request or use the Microphone.

7.2 SDKs

The client uses native iOS capabilities, sign-in/payment components, a local database, and Firebase Analytics / Crashlytics for the technical diagnostics described in Section 6.1. The Appendix identifies these SDKs, their operators, purposes, information categories, and privacy links. Diagnostics do not require access to your photo library, camera, microphone, contacts, or location; permissions for the Service’s separate features remain as listed above.

7.3 Sign-in and payments

Apple / Google sign-in and Apple IAP are subject to Apple’s and Google’s own terms and privacy policies. We sell consumable credit packs, not auto-renewable subscriptions.

8. Personal information sharing, processors, and disclosure

8.1 Personal information sharing with third-party AI models

When you request an AI feature and provide the required consent, our server transmits the information needed for that request to the relevant providers below over an encrypted connection. Profile photos may reveal your face, body appearance and background. Garment or reference photos may also contain people and other personal information. Submit only relevant content you have the right to use. “Sharing” in this section includes processing by a service provider on our behalf; it does not mean publishing your information to other users or sharing for cross-context behavioral advertising.

Transmission happens only when the feature is enabled, you initiate the request and the required consent is in place. The AI consent notice identifies the providers enabled for your request. A different model provider will not be added to your processing without updated disclosure and the required consent.

We do not send account passwords, verification codes, payment credentials or your address book as model request parameters. Phone numbers, email addresses and precise location are not structured parameters in these model requests. However, information you include in a submitted photo or text is transmitted as part of that input, so avoid unnecessary personal information. Providers may process connection information such as our server’s network address, request times and technical operation data.

We request separate AI processing consent before the first transfer of personal information. You may decline or withdraw consent by signing in to the App, opening Privacy Policy, and selecting “Withdraw AI processing consent”. This stops new related transfers; browsing and manual wardrobe management remain available. Processing already started cannot be undone. We update the notice and obtain renewed consent where required when recipients, purposes or scope change.

Provider processing and retention follow the applicable service terms, data-processing arrangements and actual settings. Our 6-month audit retention period does not define the provider’s retention period, and we do not promise zero retention by third parties. This sharing does not grant us permission to train models on your inputs or outputs. Contact the privacy email in this Policy for assistance with access or deletion requests.

8.2 Processors (service providers)

We use processors under instructions and contracts (including DPAs where required), including:

• Cloud: AWS (us-west-2);
• AI processing: Qwen, BytePlus and TypeSafe JEV (Section 8.1);
• Weather: Open-Meteo;
• Sign-in / payments: Apple, Google;
• Transactional email: Resend (Plus Five Five, Inc.) receives your email address, verification code or password-reset link, email content and delivery information to send registration and password-reset messages. See its privacy policy. Feedback is processed by our own service.

8.3 Sale / share (CCPA/CPRA)

We do not sell personal information.
We do not “share” personal information for cross-context behavioral advertising as those terms are defined under CCPA/CPRA. If that ever changes, we will update this Policy and provide an opt-out.

We do not knowingly sell or share personal information of consumers under 18.

8.4 Other disclosures

We may disclose information: with your consent; to processors as needed to run the Service; to comply with law or lawful requests; to protect life, security, or rights; or in a merger/acquisition where the successor remains bound by this Policy or new consent is obtained.

Generation audits are not made public or shared with unrelated third parties. Necessary storage processors operate within their assigned scope; lawful disclosure requirements may apply. Data sent to model vendors is limited to what is needed for the current inference request.

8.5 Public disclosure

We do not publicly disclose your personal information by default. Style Help results are visible to participants, not the general public.

9. International transfers

Account and business data are stored in the United States (AWS us-west-2). The AI and email services in Section 8 and enabled Firebase diagnostics may process request data in other countries under their service arrangements. TypeSafe describes its services as hosted in the United States. Hosting and processing locations may differ from your country of residence.

Where applicable law requires safeguards for a transfer, we must put an appropriate transfer mechanism and necessary safeguards in place before that transfer, such as applicable standard contractual clauses where required. We provide any additional notice and obtain consent where required. You may contact us at the end of this Policy for information about the safeguards applicable to your data and how to obtain a copy. This Policy does not itself establish or replace a required transfer arrangement.

10. Retention, deletion, and account closure

10.1 Retention overview

We keep personal information only as long as needed for the purposes in this Policy, unless a longer period is required by law.

If information in backups cannot immediately be deleted, it is isolated, used only for storage and necessary security protection, and cleared as backups are refreshed. Applicable deletion requests are reapplied before restored data is put back into service.

10.2 Self-service deletion

• Delete individual garments (cloud sync);
• Delete on-device generated outfit images;
• Delete Style Help results/invites;
• Reset the on-device profile.

10.3 Account deletion

Request account deletion in the App. After identity verification and completion of the closure process, associated cloud identity, sessions, wardrobe, credits, generation audit metadata and audit images are automatically deleted; failed stored-copy deletions are retried. We generally complete processing within 15 business days of receiving a valid request and necessary information, and explain progress in complex cases. Anti-abuse markers and other records follow Section 10.1. Back up images you wish to keep before closure and clear local data on other devices separately. Deletion is irreversible.

11. Your privacy rights

11.1 General

Depending on where you live, you may have rights to access, correct, delete, port, restrict, or object to certain processing, and to withdraw consent. Use in-App settings where available, or email the contact email at the end of this document and include the account details necessary to verify your request. We may need to verify your identity. We aim to respond within 15 business days, or sooner if law requires (e.g., GDPR one-month baseline with permitted extensions).

11.2 EEA / UK / Switzerland

Where applicable, you may: access and obtain a copy; rectify; erase in cases provided by law; restrict or object (including to legitimate-interests processing); data portability for data you provided that we process by contract or consent; withdraw consent at any time; lodge a complaint with your local supervisory authority.

We do not use solely automated decision-making that produces legal or similarly significant effects about you; styling suggestions and image generation are assistive tools you can choose not to use.

11.3 California (CCPA/CPRA) and similar U.S. state laws

If you are a California resident (and similarly under other state laws where applicable), you may have the right to: know categories and specific pieces collected, sources, purposes, and categories of recipients; delete (subject to exceptions); correct; opt out of sale or sharing—we do not sell or share as defined above; limit use/disclosure of sensitive personal information—to the extent we collect it, we use it to provide the Service you request; non-discrimination for exercising privacy rights.

Authorized agents may submit requests with proof of authorization.

11.4 Appeals and complaints

If you are unsatisfied with our response, reply to the contact email at the end of this document to escalate. EEA/UK users may contact their supervisory authority. U.S. users may contact their state Attorney General.

12. Security

We use reasonable technical and organizational measures, including TLS in transit; hashed passwords and session tokens; server-only secrets; access control and least privilege; on-device profile storage by default; memory-forward of original AI inputs and purpose-specific image retention limits; separated admin systems with audit logs; and staff confidentiality practices.

No system is 100% secure. If a personal-data breach occurs, we will notify you and regulators as required by law and take steps to mitigate harm.

13. Children

13.1 The Service is available only to people aged 18 or older. People under 18 must not register, use the Service or provide personal information. Stricter local eligibility requirements also apply.

13.2 If we discover that an ineligible user has provided personal information, we will restrict the account and delete the information, except where retention is legally required.

13.3 Guardians who discover that a person under 18 has used the Service or provided information may contact us using the email at the end of this Policy to request review, cessation or deletion. Acceptance does not mean we have verified age.

14. Policy updates

We may update this Policy, change the date above and notify you through the App or other reasonable means. Material changes will be highlighted. Where new consent is required, continued use does not substitute for that consent. You may decline the affected processing, stop the affected feature and still exercise your privacy rights or request account deletion.

Appendix: Processors / third parties

Firebase technical diagnostics are separately disclosed below.

Third-party SDKs

SDK names: Google Analytics for Firebase (Firebase Analytics) / Firebase Crashlytics

Third-party company: Google LLC

Function: App performance and stability diagnostics; crash-cause analysis and verification of fixes. Firebase Analytics reports the predefined technical events and elapsed times described in Section 6.1; Crashlytics reports crash traces and related diagnostic context. We have not integrated the separate Firebase Performance Monitoring SDK.

Permissions: Network access; no camera, photo-library, microphone, contacts, or location permission requested for diagnostics.

Usage scenario and information: When diagnostics are enabled, app sessions and the technical operation stages described in Section 6.1, or crashes and subsequent report delivery. Data includes the technical categories disclosed there. Reports are not associated by us with your Jemly account and do not include private content as custom diagnostic parameters. Crashlytics may use the predefined technical events as breadcrumbs to help explain a crash.

Processing and retention: Google may process technical diagnostic data on its global infrastructure, including outside your country. Crashlytics retains crash traces and associated identifiers for 90 days before starting removal from live and backup systems, as described in its provider documentation. Analytics event/identifier retention follows the applicable property settings and Google’s service rules; aggregate reports may follow different retention rules. Account deletion does not automatically identify or erase diagnostics that we have not linked to that account. For privacy requests, use the contact in Section 15; do not send passwords or private photos.

Privacy information: Firebase privacy and security, Google Privacy Policy.

SDK name: Google Sign-In for iOS

Third-party company: Google LLC

Function: Google account authentication

Permissions: Network access and the account information you authorize on the Google sign-in screen; no camera, microphone, contacts or location permission requested by this integration

Usage scenario: When you choose Google sign-in or link a Google account. The sign-in integration includes AppAuth, GTMAppAuth, GTMSessionFetcher, AppCheckCore, GoogleUtilities and FBLPromises as supporting dependencies for authorization, network requests, app verification and asynchronous operations. They are not separate chat, advertising or analytics features.

Information: Google account identifier, authorized email/basic profile, authentication tokens and technical connection information needed for sign-in.

Privacy policy: Google Privacy Policy

SDK name: GRDB.swift

Third-party company: Not applicable (open-source project maintained by Gwendal Roué and contributors)

Function: On-device SQLite data storage

Permissions: Reads and writes within the App sandbox; does not request camera, microphone, contacts or location access

Usage scenario: Browsing and organizing your wardrobe and saving local application records. This component does not transmit those records to its maintainers.

Project information: GRDB.swift

15. Contact us

• Operator: Jemly Studio
• Contact email: service@jemly.app